Administration of Web versus Paper Surveys: Mode Effects and Response Rates.

A survey of academic reference librarians in North Carolina provided data for an examination of differences in survey administration on paper and the World Wide Web. Research via the Internet is becoming more attractive for many researchers, but the effects of this medium upon research outcomes have been little explored. This study examined in particular sampling and mode effects, and response rates of Web surveys. The study found no sampling bias or mode effects in tests of the respondents' demographics and the content of responses. Response rates to Web surveys are not as high as traditional survey methods, and while responses are gathered more quickly, the paper instrument was not far behind. E-mail notices were more efficient for promoting the Web survey than paper notices. Traditional postal surveys still hold some advantages over Web surveys. Researchers must weigh the advantages in cost and speed to justify use of such instruments. Appendices contain paper survey, Web survey, sample cover letters, selected statistical test results, and survey content summary. (Contains 30 references.) (Author)